Permanent genomic modification has been a long sought after goal since the discovery that many human disorders are the result of genetic mutations that could, in theory, be corrected by providing the patient with a non-mutated gene. Permanent alterations of the genomes of cells and tissues would also be valuable for research applications, commercial products, protein production, and medical applications. Furthermore, genomic modification in the form of transgenic animals and plants has become an important approach for the analysis of gene function, the development of disease models, and the design of economically important animals and crops.
A major problem with many genomic modification methods associated with gene therapy is their lack of permanence. Life-long expression of the introduced gene is required for correction of genetic diseases. Indeed, sustained gene expression is required in most applications, yet current methods often rely on vectors that provide only a limited duration of gene expression. For example, gene expression is often curtailed by shut-off of integrated retroviruses, destruction of adenovirus-infected cells by the immune system, and degradation of introduced plasmid DNA (Anderson, W F, Nature 329:25-30, 1998; Kay, et al, Proc. Natl. Acad. Sci. USA 94:12744-12746, 1997; Verma and Somia, Nature 389:239-242, 1997). Even in shorter-term applications, such as therapy designed to kill tumor cells or discourage regrowth of endothelial tissue after restenosis surgery, the short lifetime of gene expression of current methods often limits the usefulness of the technique.
One method for creating permanent genomic modification is to employ a strategy whereby the introduced DNA becomes part of (i.e., integrated into) the existing chromosomes. Of existing methods, only retroviruses provide for efficient integration. Retroviral integration is random, however, thus the added gene sequences can integrate in the middle of another gene, or into a region in which the added gene sequence is inactive. In addition, a different insertion is created in each target cell. This situation creates safety concerns and produces an undesirable loss of control over the procedure.
Adeno-associated virus (AAV) often integrates at a specific region in the human genome. However, vectors derived from AAV do not integrate site-specifically due to deletion of the toxic rep gene (Flotte and Carter, Gene Therapy 2:357-362, 1995; Muzyczk, Curr. Topics Microbiol. Immunol. 158:97-129, 1992). The small percentage of the AAV vector population that eventually integrates does so randomly. Other methods for genomic modification include transfection of DNA using calcium phosphate co-precipitation, electroporation, lipofection, microinjection, protoplast fusion, particle bombardment, or the Ti plasmid (for plants). All of these methods produce random integration at low frequency. Homologous recombination produces site-specific integration, but the frequency of such integration is very low.
Another method that has been considered for the integration of heterologous nucleic acid fragments into a chromosome is the use of a site-specific recombinase (an example using Cre is described below). Site-specific recombinases catalyze the insertion or excision of nucleic acid fragments. These enzymes recognize relatively short, unique nucleic acid sequences that serve for both recognition and recombination. Examples include Cre (Sternberg and Hamilton, J Mol Biol 150:467-486, 1981), Flp (Broach, et al, cell-29:227-234, 1982) and R (Matsuzaki, et al, J Bacteriology 172:610-618, 1990).
One of the most widely studied site-specific recombinases is the enzyme Cre from the bacteriophage P1. Cre recombines DNA at a 34 basepair sequence called loxP, which consists of two thirteen basepair palindromic sequences flanking an eight basepair core sequence. Cre can direct site-specific integration of a loxP-containing targeting vector to a chromosomally placed loxP target in both yeast and mammalian cells (Sauer and Henderson, New Biol 2:441-449, 1990). Use of this strategy for genomic modification, however, requires that a chromosome first be modified to contain a loxP site (because this sequence is not known to occur naturally in any organism but P1 bacteriophage), a procedure which suffers from low frequency and unpredictability as discussed above. Furthermore, the net integration frequency is low due to the competing excision reaction also mediated by Cre. Similar concerns arise in the conventional use of other, well-known, site-specific recombinases.
A need still exists, therefore, for a convenient means by which chromosomes can be permanently modified in a site-specific manner. The present invention addresses that need.